As a result of advances in automated sequencing technology, much work has been carried out on determining coding regions of DNA resulting in the full sequencing of many animal genomes including the human genome. It has been realised for many years that the majority of genomic DNA, however, is non-coding and this material was once considered as “junk” DNA. Analysis of the non-coding regions of DNA is now being considered as important in the study of gene expression and function. Methylation states or patterns in nucleic acid, particularly genomic DNA, is thought to have a functional or regulatory role in gene expression and control in animals.
It has been demonstrated that, in single stranded DNA, sodium bisulphite preferentially deaminates cytosine to uracil, compared to a very slow rate of deamination of 5-methylcytosine to thymine (Shapiro, R., DiFate, V., and Welcher, M, (1974) J. Am. Chem. Soc. 96: 906-912). This observation served as the basis for the development of the bisulphite genomic sequencing protocol of Frommer et al 1992 [Frommer M, McDonald L E, Millar D S, Collis C M, Watt F, Grigg G W, Molloy P L and Paul C L. PNAS 89: 1827-1831 (1992), which is incorporated herein by reference]. In summary, this method as presently practiced involves the following general steps: alkaline denaturation of DNA; deamination using sodium bisulphite; desulphonation by desalting followed by sodium hydroxide treatment; neutralization and desalting.
One of the major disadvantages of the bisulphite modification procedure and the established variation thereof is that it has been shown that the procedure results in the degradation of between 84-96% of the original input DNA (Grunau et al. Nucleic Acids Research 29 (13) e65;(2001). The high loss associated with the procedure means that practically it is very difficult to successfully analyse small numbers of cells for their methylation status, or successfully analyse ancient archival specimens in which the DNA is already in a partially degraded state. In addition, due to inherent degradation of the current methods, it is not possible to sequence and assemble the complete genome of an organism to determine its genome-wide methylation profile in the same manner as has been successfully applied by the public Human Genome Project (International Human Genome Sequencing Consortium, 2001, Nature, 409, 860-921) or the private CELERA sequencing project (J Craig Venter et al., 2001, Science, 291, 1304-1351) as the DNA would be so fragmented as it would not be able to be cloned, sequenced, and assembled in any meaningful way owing to the huge number of “gaps” in the sequence.
A further disadvantage with the bisulphite method as presently practiced is that, in general, only small fragments of DNA can be amplified. Experience shows that generally less than about 500 base pairs (bp) can be successfully treated and amplified. The present technique is not applicable to new molecular biological methods such as Long Distance polymerase chain reaction (PCR) which has made it possible to amplify large regions of untreated genomic DNA, generally up to about 50 kb. At present, it is not even possible to analyse the methylation status of intact genes, as a large number of genes in mammalian genomes exceed 50 kb in length.
To look at the methylation status of even relatively small genes (<4 kb), PCR reactions have had to be staggered across the gene region of interest (D. S Millar, K. K Ow, C. L. Paul, P. J. Russell, P. L. Molloy, S. J. Clark, 1999, Oncogene, 18(6):1313-24; Millar D S, Paul C L, Molloy P L, Clark S J. (2000). J Biol Chem; 275(32):24893-9). The methods presently used for bisulphite DNA treatment have also been laborious and time consuming. Standard methods typically require multiple tube changes, column purifications, dialysis, embedding the DNA in agarose beads or the addition of additives to the reaction in an attempt to reduce problems such as non-conversion of certain regions of genomic DNA. Thus a more reliable method that does not lead to substantial DNA degradation, and which overcomes or at least reduces the multiple problems associated with present DNA treatment, is required.